The question of which ganglion cells send color signals to the brain is unresolved. The M and P construct provides that achromatic information is encoded by parasol cells, and chromatic information by midget cells that project, respectively, to the M and P layers of the LGN. However, K layers in the LGN, ventral to each M and P layer, project to the cytochrome oxidase-rich blue/yellow and red/green color blobs in layers 2 and 3 of VI, and it is known in the case of the blue/yellow pathway, that there are blue/yellow small bistratified ganglion cells (i.e., not midget cells) that project to the LGN K cells. The central hypothesis of the present proposal is that ganglion cells that innervate other K cells in the LGN (i.e., alpha-CamK-positive gamma and epsilon cells) (1) are sufficient to provide additional vision channels, and (2) include ganglion cells that pools spatially co-extensive signals from M and L cones via an opponent bipolar cell to underlie red/green color vision. The specific aims to test this hypothesis in macaque retina are: (1) to measure the spatial distribution of gamma and epsilon cells using alpha-CamK-staining of the retina; (2) to establish their presynaptic circuitry via electron microscopy of serial sections through stained and injected cells; and (3) to determine the localization of glutamate receptors to bipolar cell dendrites via electron microscopy of serial sections through stained and injected cells.